In this paper, we show how to create paraphrastic sentence embeddings using the Paraphrase Database (Ganitkevitch et al., 2013), an extensive semantic resource with millions of phrase pairs. We consider several compositional architectures and evaluate them on 24 textual similarity datasets encompassing domains such as news, tweets, web forums, news headlines, machine translation output, glosses, and image and video captions. We present the interesting result that simple compositional architectures based on updated vector averaging vastly outperform long short-term memory (LSTM) recurrent neural networks and that these simpler architectures allow us to learn models with superior generalization. Our models are efficient, very easy to use, and competitive with task-tuned systems. We make them available to the research community with the hope that they can serve as the new baseline for further work on universal paraphrastic sentence embeddings.

研究使用Paraphrase Database监督学习通用的、同义句的句子嵌入，比较6种组成结构并在相似度数据集和三个NLP任务上进行评估，在领域内复杂的结构（如LSTM）有更好的性能，但在领域外使用简单架构如词平均比LSTM表现更好，同时演示如何将预训练的句子嵌入与监督任务相结合。

通往通用释义句子嵌入的路途